Re: 8.1.5 Partition Row Movement performance

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sun, 26 Sep 1999 19:33:59 +0100
Message-ID: <938371252.10177.0.nnrp-13.9e984b29@news.demon.co.uk>

Icant' seem to find such an extreme difference. The current gets in the partitioned case seem to be slightly higher, and the redo log generated seems to be about 20%higher on a few quick tests.

Can you give a more detailed description of the test. Are you sure that the reads to select the row to be updated were doing partition elimination ? That could explain a doubling in cost.

Jonathan Lewis
Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk

Simon Quinn wrote in message ...
>I've just been testing the performance of the new 8.1.5 feature, Row
>Movement between partitions of a table. My tests were simply to tests
>the difference in speed between inserting into a target table and
>deleting from the source compared to just updating a partitioned table
>to force a row to move from one partition to another.
>
>From my tests it takes more than twice as long to do it via row
>movement. Anybody have experience of this?
Received on Sun Sep 26 1999 - 13:33:59 CDT